Recently, individual differences that might predict risk for development of dependence and/or risk for relapse have been identified and are being investigated. Of interest to this component are individual differences in "trait-impulsivity." Findings in animals and humans suggest that a dysfunctional brain dopamine system underlies impulsive traits. Given that the brain dopamine system underlies both impulsivity and reward areas in the brain, it is not surprising that individuals high in trait-impulsivity show different brain regional activity to reward than individuals low in trait-impulsivity. Interestingly, these areas of activation are similar to what has been observed in non-treatment seeking alcoholics (NTSA) following presentation of alcohol cues. We found that aripiprazole, a partial dopamine receptor agonist that putatively "stabilizes" the dopamine system, reduces drinking and blocks alcohol cue-induced ventral striatal activity in NTSAs, and it does so most clearly in individuals with high trait-impulsivity. The aim of the current proposal is to replicate and extend this original finding with aripiprazole, in a 2 medication group x 2 trait-impulsivity group prospective experimental design. In addition to studying the interaction of aripiprazole and trait-impulsivity, we also intend to systematically evaluate the role of trait-impulsivity and aripiprazole, per se, as main effects on drinking and craving using our well-validated, natural drinking, brain imaging, and bar-lab alcohol consumption paradigm. Finally, we intend to explore whether functional genetic variants in the brain dopamine system might underlie trait-impulsivity effects and/or aripiprazole response. We will screen about 120 NTSAs and randomize 100 to receive aripiprazole (N=50) (up to 15mg) or matching placebo (N=50) for 8 days. Barratt Impulsiveness Scale (BIS) scores will be used to divide the groups into low versus high trait impulsivity and will be an urn variable to insure equal distribution into both medication groups. After 6-days of natural drinking observation, all subjects will undergo an alcohol cue-induced brain activation (fMRI) paradigm (day-7) and the next day (day-8) will participate in a drinking bar-lab experiment. On day-9, subjects will receive educational counseling about heavy drinking and receive payment for participation.